 ************Do-file that reproduces statistical analyses in the online appendix
 ***********Paper, Cordova&Hiskey (2019), "Development Context and the Political Behavior of Remittance Recipients in Latin America and the Caribbean"
 *****Run these models using the 2008 LAPOP dataset provided
 **Table A4. Effect of Remittance Dependence on Life Satisfaction among Recipients by Level of Development as Measured by the Human Development Index
 set more off
*Model 1
meologit ls3r i.q10brr eff1 pol1r vic1extrr aoj11r edr	 mujer	q2	agesquared	urr	q10 if	year==2010 & q10br>0 & hdi_quintiles==1
*Model 2
meologit ls3r i.q10brr eff1 pol1r vic1extrr aoj11r edr	 mujer	q2	agesquared	urr	q10 if	year==2010 & q10br>0 & hdi_quintiles==2
*Model 3
meologit ls3r i.q10brr eff1 pol1r vic1extrr aoj11r edr	 mujer	q2	agesquared	urr	q10 if	year==2010 & q10br>0 & hdi_quintiles==3
*Model 4
meologit ls3r i.q10brr eff1 pol1r vic1extrr aoj11r edr	 mujer	q2	agesquared	urr	q10 if	year==2010 & q10br>0 & hdi_quintiles==4
*Model 5
meologit ls3r i.q10brr eff1 pol1r vic1extrr aoj11r edr	 mujer	q2	agesquared	urr	q10 if	year==2010 & q10br>0 & hdi_quintiles==5

**Figure A1. Effect of Remittance Dependence on Life Satisfaction among Recipients in the Poorest Countries in Latin America and Caribbean
**(Results based on Model 1 in Table A4)
**Panel A
 margins, over(q10brr) expression(predict(mu fixedonly outcome(4)))
    #delimit; 
 marginsplot, recast(line) noci 
 plot1opts(lwidth(medthick)lcolor(black))
ytit("Mean Predicted Probability") xtit("Remittance Dependence")
 title("Panel A", size(medsmall))
ylabel(, nogrid angle(horizontal)) plotregion(color(white)) graphregion(color(white));
 **Panel B
 #delimit cr
 margins, over(q10brr) contrast(overjoint effect) expression(predict(mu fixedonly outcome(4)))
 marginsplot
 
 ********Table A5. Effect of Remittance Dependence on Community and Political Participation across Countries with Different Development Levels
 ***Community Participation
 melogit cp8rr c.q10br##c.hdi2010  gdpgrowth09 eff1 pol1r vic1extrr aoj11r edr	 mujer	q2	agesquared	urr	q10 if	year==2010	|| pais:	, cov(uns)
 **Figure A2
  margins, dydx(q10br) at(hdi2010=(0.45 (.05) 0.80)) predict(mu fixedonly) 
 #delimit; 
 marginsplot, recast(line) recastci(rarea) ciopts(fcolor(gs14) color(gs14)) 
 plot1opts(lcolor(gs0) lpattern(dash)lwidth(thin)) 
 ylabel(, nogrid)plotregion(color(white)) graphregion(color(white));
 #delimit cr
 ***Requested Help from Local Government
 melogit cp4ar c.q10br##c.hdi2010  gdpgrowth09 eff1 pol1r vic1extrr aoj11r edr	 mujer	q2	agesquared	urr	q10 if	year==2010	|| pais:	, cov(uns)
 
 **Figure A3
  margins, dydx(q10br) at(hdi2010=(0.45 (.05) 0.80)) predict(mu fixedonly) 
 #delimit; 
 marginsplot, recast(line) recastci(rarea) ciopts(fcolor(gs14) color(gs14)) 
 plot1opts(lcolor(gs0) lpattern(dash)lwidth(thin)) 
 ylabel(, nogrid)plotregion(color(white)) graphregion(color(white));
 #delimit cr
  **Voting
  melogit vb2r c.q10br##c.hdi2010  gdpgrowth09 eff1 pol1r vic1extrr aoj11r edr	 mujer	q2	agesquared	urr	q10 if	year==2010	|| pais:	, cov(uns)
  
  **Figure A4
  margins, dydx(q10br) at(hdi2010=(0.45 (.05) 0.80)) predict(mu fixedonly) 
 #delimit; 
 marginsplot, recast(line) recastci(rarea) ciopts(fcolor(gs14) color(gs14)) 
 plot1opts(lcolor(gs0) lpattern(dash)lwidth(thin)) 
 ylabel(, nogrid)plotregion(color(white)) graphregion(color(white));
 #delimit cr
 
****Table A6.Participation Models Controlling for Migration Intentions 
 
 ***Community Participation, controlling for migration intentions (q14)
set more off
meologit cp8r i.q10ar##c.hdi2010  gdpgrowth09 q14r eff1 pol1r vic1extrr aoj11r edr	 mujer	q2	agesquared	urr	q10 if	year==2010	|| pais:	, cov(uns)

***Figure A5 
margins, over(q10ar) at(hdi2010=(0.45 (.05) 0.80)) expression(predict(mu fixedonly outcome(66))+ predict(mu fixedonly outcome(100))) 
 #delimit; 
 marginsplot, recast(line) noci  
ylabel(, nogrid)plotregion(color(white)) graphregion(color(white));
 #delimit cr

 margins , over(q10ar) at(hdi2010=(0.45 (.05) 0.80)) contrast(overjoint effect) expression(predict(mu fixedonly outcome(66))+ predict(mu fixedonly outcome(100)))
 #delimit; 
 marginsplot, recast(line) recastci(rarea) ciopts(fcolor(gs14) color(gs14)) 
 ylabel(, nogrid)plotregion(color(white)) graphregion(color(white));
 #delimit cr

 ***Requested help from local government, controlling for migration intentions (q14)
 melogit cp4ar i.q10ar##c.hdi2010  gdpgrowth09 q14r eff1 pol1r vic1extrr aoj11r edr	 mujer	q2	agesquared	urr	q10 if	year==2010	|| pais:	, cov(uns)

margins, over(q10ar) at(hdi2010=(0.45 (.05) 0.80)) predict(mu fixedonly ) 
 #delimit; 
 marginsplot, recast(line) noci  
ylabel(, nogrid)plotregion(color(white)) graphregion(color(white));
 #delimit cr

margins , over(q10ar) at(hdi2010=(0.45 (.05) 0.80)) contrast(overjoint effect) predict(mu fixedonly )
 #delimit; 
 marginsplot, recast(line) recastci(rarea) ciopts(fcolor(gs14) color(gs14)) 
 ylabel(, nogrid)plotregion(color(white)) graphregion(color(white));
 #delimit cr

 ****Voting, controlling for migration intentions (q14)
set more off
melogit vb2r i.q10ar##c.hdi2010  gdpgrowth09 q14r eff1 pol1r vic1extrr aoj11r edr	 mujer	q2	agesquared	urr	q10 if	year==2010	|| pais:	, cov(uns)

margins, over(q10ar) at(hdi2010=(0.45 (.05) 0.80)) predict(mu fixedonly ) 
 #delimit; 
 marginsplot, recast(line) noci  
ylabel(, nogrid)plotregion(color(white)) graphregion(color(white));
 #delimit cr

margins , over(q10ar) at(hdi2010=(0.45 (.05) 0.80)) contrast(overjoint effect) predict(mu fixedonly )
 #delimit; 
 marginsplot, recast(line) recastci(rarea) ciopts(fcolor(gs14) color(gs14)) 
 ylabel(, nogrid)plotregion(color(white)) graphregion(color(white));
 #delimit cr

***Table A7. Participation Models Controlling for Freedom House (partly free pf=0; free pf=1)
**Community participation
set more off
meologit cp8r i.q10ar##c.hdi2010 pf gdpgrowth09  eff1 pol1r vic1extrr aoj11r edr	 mujer	q2	agesquared	urr	q10 if	year==2010	|| pais:	, cov(uns)

margins, over(q10ar) at(hdi2010=(0.45 (.05) 0.80)) expression(predict(mu fixedonly outcome(66))+ predict(mu fixedonly outcome(100))) 
 #delimit; 
 marginsplot, recast(line) noci  
ylabel(, nogrid)plotregion(color(white)) graphregion(color(white));
 #delimit cr


set more off
  margins , over(q10ar) at(hdi2010=(0.45 (.05) 0.80)) contrast(overjoint effect) expression(predict(mu fixedonly outcome(66))+ predict(mu fixedonly outcome(100))) 
 #delimit; 
 marginsplot, recast(line) recastci(rarea) ciopts(fcolor(gs14) color(gs14)) 
 ylabel(, nogrid)plotregion(color(white)) graphregion(color(white));
 #delimit cr


**Requested help from local government
set more off
melogit cp4ar i.q10ar##c.hdi2010  gdpgrowth09 pf eff1 pol1r vic1extrr aoj11r edr	 mujer	q2	agesquared	urr	q10 if	year==2010	|| pais:	, cov(uns)

margins, over(q10ar) at(hdi2010=(0.45 (.05) 0.80)) predict(mu fixedonly ) 
 #delimit; 
 marginsplot, recast(line) noci  
ylabel(, nogrid)plotregion(color(white)) graphregion(color(white));
 #delimit cr

 
 set more off
margins , over(q10ar) at(hdi2010=(0.45 (.05) 0.80)) contrast(overjoint effect) predict(mu fixedonly )
 #delimit; 
 marginsplot, recast(line) recastci(rarea) ciopts(fcolor(gs14) color(gs14)) 
 ylabel(, nogrid)plotregion(color(white)) graphregion(color(white));
 #delimit cr

**Voting
melogit vb2r i.q10ar##c.hdi2010 pf gdpgrowth09  eff1 pol1r vic1extrr aoj11r edr	 mujer	q2	agesquared	urr	q10 if	year==2010	|| pais:	, cov(uns)

margins, over(q10ar) at(hdi2010=(0.45 (.05) 0.80)) predict(mu fixedonly ) 
 #delimit; 
 marginsplot, recast(line) noci  
ylabel(, nogrid)plotregion(color(white)) graphregion(color(white));
 #delimit cr


set more off
margins , over(q10ar) at(hdi2010=(0.45 (.05) 0.80)) contrast(overjoint effect) predict(mu fixedonly )
 #delimit; 
 marginsplot, recast(line) recastci(rarea) ciopts(fcolor(gs14) color(gs14)) 
 ylabel(, nogrid)plotregion(color(white)) graphregion(color(white));
 #delimit cr


****Table A8. Participation Models Controlling for Polity V
set more off
meologit cp8r i.q10ar##c.hdi2010 gdpgrowth09 polity_2009 eff1 pol1r vic1extrr aoj11r edr	 mujer	q2	agesquared	urr	q10 if	year==2010	|| pais:	, cov(uns)

margins, over(q10ar) at(hdi2010=(0.45 (.05) 0.80)) expression(predict(mu fixedonly outcome(66))+ predict(mu fixedonly outcome(100))) 
 #delimit; 
 marginsplot, recast(line) noci  
ylabel(, nogrid)plotregion(color(white)) graphregion(color(white));
 #delimit cr

set more off
  margins , over(q10ar) at(hdi2010=(0.45 (.05) 0.80)) contrast(overjoint effect) expression(predict(mu fixedonly outcome(66))+ predict(mu fixedonly outcome(100))) 
 #delimit; 
 marginsplot, recast(line) recastci(rarea) ciopts(fcolor(gs14) color(gs14)) 
 ylabel(, nogrid)plotregion(color(white)) graphregion(color(white));
 #delimit cr

 ****Requested help from local gov
 melogit cp4ar i.q10ar##c.hdi2010  gdpgrowth09 polity_2009 eff1 pol1r vic1extrr aoj11r edr	 mujer	q2	agesquared	urr	q10 if	year==2010	|| pais:	, cov(uns)

margins, over(q10ar) at(hdi2010=(0.45 (.05) 0.80)) predict(mu fixedonly ) 
 #delimit; 
 marginsplot, recast(line) noci  
ylabel(, nogrid)plotregion(color(white)) graphregion(color(white));
 #delimit cr
 
 set more off
margins , over(q10ar) at(hdi2010=(0.45 (.05) 0.80)) contrast(overjoint effect) predict(mu fixedonly )
 #delimit; 
 marginsplot, recast(line) recastci(rarea) ciopts(fcolor(gs14) color(gs14)) 
 ylabel(, nogrid)plotregion(color(white)) graphregion(color(white));
 #delimit cr

 ***Voting
 set more off
melogit vb2r i.q10ar##c.hdi2010 gdpgrowth09 polity_2009 eff1 pol1r vic1extrr aoj11r edr	 mujer	q2	agesquared	urr	q10 if	year==2010	|| pais:	, cov(uns)
outreg using VB_PF.doc, se varlabels starlevels(10 5 1 0.1)starloc(1) sigsymbol(+, *,**, ***)
margins, over(q10ar) at(hdi2010=(0.45 (.05) 0.80)) predict(mu fixedonly ) 
 #delimit; 
 marginsplot, recast(line) noci  
ylabel(, nogrid)plotregion(color(white)) graphregion(color(white));
 #delimit cr

set more off
margins , over(q10ar) at(hdi2010=(0.45 (.05) 0.80)) contrast(overjoint effect) predict(mu fixedonly )
 #delimit; 
 marginsplot, recast(line) recastci(rarea) ciopts(fcolor(gs14) color(gs14)) 
 ylabel(, nogrid)plotregion(color(white)) graphregion(color(white));
 #delimit cr

***Table A9.Endogeneity Test: Average Treatment Effect (ATE) of Receiving Remittances Estimated using the Augmented Inverse Propensity Weighted (AIPW) procedure
 teffects aipw (cp4ar    eff1 pol1r vic1extrr aoj11 edr	 mujer	q2	agesquared	urr	q10  , logit) (q10ar pn4r m1r edr q10 mujer q2 urr )


 ***(Run these models using the 2008 LAPOP dataset provided)Table A10. Effect of Remittances on Economic Perceptions across Development Contexts 
  ***Model 1
estimates clear
melogit idio1rr i.q10ar   gdpgrowth08 m1r edr	 mujer	q2	urr	q10 if year==2008	|| pais:	, cov(uns) 
***Model 2
melogit idio1rr i.q10ar c.hdi2008  gdpgrowth08 m1r edr	 mujer	q2	urr	q10 if year==2008	|| pais:	, cov(uns) 
***Model 3
melogit idio1rr i.q10ar##c.hdi2008  gdpgrowth08 m1r edr	 mujer	q2	urr	q10 if year==2008	|| pais:	, cov(uns)  


***Figure A11. Perceptions of the Personal and National Economy (Panels A and B)

margins, over(q10ar) at(hdi2008=(0.45 (.05) 0.80)) predict(mu fixedonly)
  #delimit; 
 marginsplot, recast(line) noci 
 plot2opts(lwidth(medthick)lcolor(black)) plot1opts(lwidth(medthick) lcolor(black) lpattern(dash))
ytit("Mean Predicted Probability") xtit(""Development Level""(Human Development Index)"")
 title("Panel A", size(medsmall))
 legend(rows(1) size(small) order(1 2)  label(2 "Remittance Recipient") label(1 "Non-Remittance Recipient")) 
ylabel(, nogrid angle(horizontal)) plotregion(color(white)) graphregion(color(white));
 #delimit cr

 margins, over(q10ar) at(hdi2008=(0.45 (.05) 0.80)) contrast(overjoint effect) predict(mu fixedonly)
  #delimit; 
 marginsplot, recast(line) recastci(rarea) ciopts(fcolor(gs14) color(gs14)) 
 plot1opts(lcolor(gs0) lwidth(medthick))
 yline(0, lcolor(red) lwidth(medthick))
 ytit("Difference in Mean Predicted Probability") xtit("Development Level""(Human Development Index)")
  legend(on) legend(rows(1) size(small) order(1 ) label(1 "95% Confidence Intervals")) 
 title("Panel B" " " , size(medsmall))
 title("Difference in Mean Predicted Probability between" "Remittances Recipients and Non-Recipients", suffix size(medsmall))
  ylabel(, nogrid angle(horizontal)) plotregion(color(white)) graphregion(color(white));
  #delimit cr
 

***(Run these models using the 2008 LAPOP dataset)Table A10. Effect of Remittances on Economic Perceptions across Development Contexts 
***Figure A11. Perceptions of the Personal and National Economy 
***Model 4
melogit soct1rr i.q10ar   gdpgrowth08  m1r  edr	 mujer	q2	urr	q10 if year==2008	|| pais:	, cov(uns) 

***Model 5
melogit soct1rr i.q10ar  c.hdi2008 gdpgrowth08  m1r  edr	 mujer	q2	urr	q10 if year==2008	|| pais:	, cov(uns) 

***Model 6
melogit soct1rr i.q10ar##c.hdi2008  gdpgrowth08  m1r  edr	 mujer	q2	urr	q10 if year==2008	|| pais:	, cov(uns) 

****Figure A11. Perceptions of the Personal and National Economy (Panels C and D)
estimates restore Model3
  margins, over(q10ar) at(hdi2008=(0.45 (.05) 0.80)) predict(mu fixedonly)
  #delimit; 
 marginsplot, recast(line) noci 
 plot2opts(lwidth(medthick)lcolor(black)) plot1opts(lwidth(medthick) lcolor(black) lpattern(dash))
ytit("Mean Predicted Probability") xtit(""Development Level""(Human Development Index)"")
 title("Panel C", size(medsmall))
 legend(rows(1) size(small) order(1 2)  label(2 "Remittance Recipient") label(1 "Non-Remittance Recipient")) 
ylabel(, nogrid angle(horizontal)) plotregion(color(white)) graphregion(color(white));
 #delimit cr
 
  margins , over(q10ar) at(hdi2008=(0.45 (.05) 0.80)) contrast(overjoint effect) predict(mu fixedonly)
 #delimit; 
 marginsplot, recast(line) recastci(rarea) ciopts(fcolor(gs14) color(gs14)) 
 plot1opts(lcolor(gs0) lwidth(medthick))
 yline(0, lcolor(red) lwidth(medthick))
 ytit("Difference in Mean Predicted Probability") xtit("Development Level""(Human Development Index)")
  legend(on) legend(rows(1) size(small) order(1 ) label(1 "95% Confidence Intervals")) 
 title("Panel D" " " , size(medsmall))
 title("Difference in Mean Predicted Probability between" "Remittances Recipients and Non-Recipients", suffix size(medsmall))
  ylabel(, nogrid angle(horizontal)) plotregion(color(white)) graphregion(color(white));
 #delimit cr

***(Run these models using the 2008-2014 merged LAPOP dataset provided)
****Table A11. Civic and Political Participation at the Local Level (based on data from the 2008, 2010, 2012, and 2014 LAPOP surveys)
 set more off
***Model 1
meologit cp8r i.q10ar    GDPgr eff1 pol1r vic1extr aoj11r edr	 mujer	q2	agesquared	urr	q10 	|| pais:	, cov(uns) || year: 
***Model 2
meologit cp8r i.q10ar c.HDI   GDPgr eff1 pol1r vic1extr aoj11r edr	 mujer	q2	agesquared	urr	q10 	|| pais:	, cov(uns) || year: 
***Model 3
meologit cp8r i.q10ar##c.HDI   GDPgr eff1 pol1r vic1extr aoj11r edr	 mujer	q2	agesquared	urr	q10 	|| pais:	, cov(uns) || year: 

*********Figure A12. Effect of Remittances on Community and Local Political Participation (Panels A and B)
margins, over(q10ar) at(HDI=(0.45 (.05) 0.80)) expression(predict(mu fixedonly outcome(66))+ predict(mu fixedonly outcome(100)))
  #delimit; 
 marginsplot, recast(line) noci 
plot2opts(lwidth(medthick)lcolor(black)) plot1opts(lwidth(medthick) lcolor(black) lpattern(dash))
ytit("Mean Predicted Probability") xtit(""Development Level""(Human Development Index)"")
 title("Panel A", size(medsmall))
 legend(rows(1) size(small) order(1 2)  label(2 "Remittance Recipient") label(1 "Non-Remittance Recipient")) 
ylabel(, nogrid angle(horizontal)) plotregion(color(white)) graphregion(color(white));
 #delimit cr

  margins , over(q10ar) at(HDI=(0.45 (.05) 0.80)) contrast(overjoint effect) expression(predict(mu fixedonly outcome(66))+ predict(mu fixedonly outcome(100)))
   #delimit; 
 marginsplot, recast(line) recastci(rarea) ciopts(fcolor(gs14) color(gs14))
 plot1opts(lcolor(gs0) lwidth(medthick))
 yline(0, lcolor(red) lwidth(medthick))
 ytit("Difference in Mean Predicted Probability") xtit("Development Level""(Human Development Index)")
  legend(on) legend(rows(1) size(small) order(1 ) label(1 "95% Confidence Intervals")) 
 title("Panel B" " " , size(medsmall))
 title("Difference in Mean Predicted Probability between" "Remittances Recipients and Non-Recipients", suffix size(medsmall))
  ylabel(, nogrid angle(horizontal)) plotregion(color(white)) graphregion(color(white));
 #delimit cr

***Model 4
melogit cp4ar i.q10ar    GDPgr eff1 pol1r vic1extr aoj11r edr	 mujer	q2	agesquared	urr	q10 	|| pais:	, cov(uns) || year: 

***Model 5
melogit cp4ar i.q10ar c.HDI   GDPgr eff1 pol1r vic1extr aoj11r edr	 mujer	q2	agesquared	urr	q10 	|| pais:	, cov(uns) || year: 

***Model 6
melogit cp4ar i.q10ar##c.HDI   GDPgr eff1 pol1r vic1extr aoj11r edr	 mujer	q2	agesquared	urr	q10 	|| pais:	, cov(uns) || year: 

***Figure A12. Effect of Remittances on Community and Local Political Participation (Paneld C and D)
margins, over(q10ar) at(HDI=(0.45 (.05) 0.80)) predict(mu fixedonly ) 
 #delimit; 
 marginsplot, recast(line) noci  
 plot2opts(lwidth(medthick)lcolor(black)) plot1opts(lwidth(medthick) lcolor(black) lpattern(dash))
ytit("Mean Predicted Probability") xtit(""Development Level""(Human Development Index)"")
 title("Panel A", size(medsmall))
 legend(rows(1) size(small) order(1 2)  label(2 "Remittance Recipient") label(1 "Non-Remittance Recipient")) 
ylabel(, nogrid angle(horizontal)) plotregion(color(white)) graphregion(color(white));
 #delimit cr

margins , over(q10ar) at(HDI=(0.45 (.05) 0.80)) contrast(overjoint effect) predict(mu fixedonly )
 #delimit; 
 marginsplot, recast(line) recastci(rarea) ciopts(fcolor(gs14) color(gs14)) 
 plot1opts(lcolor(gs0) lwidth(medthick))
 yline(0, lcolor(red) lwidth(medthick))
 ytit("Difference in Mean Predicted Probability") xtit("Development Level""(Human Development Index)")
  legend(on) legend(rows(1) size(small) order(1 ) label(1 "95% Confidence Intervals")) 
  title("Panel B" " " , size(medsmall))
 title("Difference in Mean Predicted Probability between" "Remittances Recipients and Non-Recipients", suffix size(medsmall))
  ylabel(, nogrid angle(horizontal)) plotregion(color(white)) graphregion(color(white));
 #delimit cr

***Table A12. Voter Turnout in National Elections (based on data from the 2008, 2010, 2012, and 2014 LAPOP surveys)
 set more off
***Model 1
estimates clear
melogit vb2r i.q10ar  GDPgr   eff1 pol1r vic1extr aoj11r edr	 mujer	q2	agesquared	urr	q10 	|| pais:	, cov(uns) || year: 
***Model 2
melogit vb2r i.q10ar c.HDI   GDPgr eff1 pol1r vic1extr aoj11r edr	 mujer	q2	agesquared	urr	q10 	|| pais:	, cov(uns) || year: 
***Model 3
melogit vb2r i.q10ar##c.HDI   GDPgr eff1 pol1r vic1extr aoj11r edr	 mujer	q2	agesquared	urr	q10 	|| pais:	, cov(uns) || year: 

***Figure A13. Effect of Remittances on Voting Behavior
margins, over(q10ar) at(HDI=(0.45 (.05) 0.80)) predict(mu fixedonly ) 
  #delimit; 
 marginsplot, recast(line) noci 
 plot2opts(lwidth(medthick)lcolor(black)) plot1opts(lwidth(medthick) lcolor(black) lpattern(dash))
ytit("Mean Predicted Probability") xtit(""Development Level""(Human Development Index)"")
 title("Panel A", size(medsmall))
 legend(rows(1) size(small) order(1 2)  label(2 "Remittance Recipient") label(1 "Non-Remittance Recipient")) 
ylabel(, nogrid angle(horizontal)) plotregion(color(white)) graphregion(color(white));
  #delimit cr

margins , over(q10ar) at(HDI=(0.45 (.05) 0.80)) contrast(overjoint effect) predict(mu fixedonly )
 #delimit; 
 marginsplot, recast(line) recastci(rarea) ciopts(fcolor(gs14) color(gs14))
 plot1opts(lcolor(gs0) lwidth(medthick))
 yline(0, lcolor(red) lwidth(medthick))
 ytit("Difference in Mean Predicted Probability") xtit("Development Level""(Human Development Index)")
  legend(on) legend(rows(1) size(small) order(1 ) label(1 "95% Confidence Intervals")) 
 title("Panel B" " " , size(medsmall))
 title("Difference in Mean Predicted Probability between" "Remittances Recipients and Non-Recipients", suffix size(medsmall))
  ylabel(, nogrid angle(horizontal)) plotregion(color(white)) graphregion(color(white));
  #delimit cr
